A Chat with a Virtual Customer Service Rep
How far are we from virtual customer service reps — bots, if you will — that are absolutely indistinguishable from human sales clerks? Bots that could pass a
Turing test — leaving you unaware if you are conversing with a human being or a computerprogram. I got to thinking about that the other day when I came across something called NeuroServer 3.0, just released by NativeMinds Inc.
San Francisco-based NativeMinds markets a suite of products and services to create automated virtual representatives called vReps for e-business customer service, sales and marketing.
The vReps are designed to answer customer questions (typed in a dialog box)
via natural language dialog and rules-based display of a variety of
information systems, including Web search, customer relationship management (CRM) systems and other customer-facing databases.
The heart and soul (if you will allow me a little latitude here) of a vRep is
the company's NeuroServer product suite, which acts as the brains for the
vReps. The patent-pending software matches patterns, elects "best-fit"
responses, understands conversational context and selects actions
automatically from the natural language questions customers ask.
The company Web page has a vRep called Nicole available for a dialog. She
looks a little like Diana Rigg back in the days of the TV show "The Avengers."
When I asked Nicole who is the CEO of the company, she told me: "Walter
Tackett is the CEO and President of NativeMinds. Please tell me your name."
Me: "Why?"
Nicole: "I need your name for no other reason than to make our conversation
more pleasant. Please tell me your name."
Already I didn't much like her tone (maybe I spent too much time in New York
City in my formative years), but I typed in my name anyway: "Beth Cox."
Nicole: "Hi Beth! I would be happy to answer any question directly or you can
click on any of the following questions.
These were links:
Show me what a vRep is.
Show me your Services.
Can you show me your products?
Me: "Is this a public company?"
Nicole: "It's likely that we will go the IPO route, however we can't discuss
any specific plans or competitive strategies with the general public. Please
watch our Web site for updates."
Me: "How do I get to your offices from the Fisherman's Wharf area of San
Francisco?"
Nicole: "Here are the driving instructions to our Headquarters in San
Francisco." (Up popped a window with directions and a map.)
Me: "What is the price for NeuroServer 3.0?"
Nicole: "The cost of our products depends on the needs and scope of your
project. Our enterprise customers have found that the software usually pays
for itself in about twelve to sixteen months. For more information, please
email [email protected] or contact a sales rep at (800) 968 1002.
Me: Aren't you a sales rep?
Nicole: If you would like to talk to a Sales Rep or obtain information,
please email [email protected] or you can reach our Sales office
toll-free at (800) 968 1002.
Me: "Apparently not."
Nicole: "When I hit something I don't know how to respond to, I log it to a
file. This is one of those times."
Me: I see.
Nicole: "Our Professional Services group can help you create, maintain, and
implement a vRep!"
About here is where my patience ran out, but I have to admit I was impressed.
Clearly this app is not good enough yet to pass a Turing test. The test was
introduced by Alan M. Turing (1912-1954) as "the imitation game."
When talking about the Turing test today what is generally understood is the
following:
The interrogator is connected to one person and one machine via a terminal and therefore can't see her counterparts. Her task is to find out which of the
two candidates is the machine, and which is the human only by asking them
questions. If the machine can "fool" the interrogator, it is, ipso facto,
intelligent.
Clearly there are other measures of intelligence and this test has been
subject to different kinds of criticism and has been at the heart of many
discussions in AI (artifical intelligence), philosophy and cognitive science for the past 50 years.
I asked Jody Ferrara at NativeMinds if their app has ever passed a
Turing test. Here's what he e-mailed me:
"Good question. You are actually the first to ask me about the Turing test.
NativeMinds vReps are designed to be ‘almost human' and very conversational,
without trying to fool anyone that they are indistinguishable from real
people…
The companies that are using NativeMinds for customer service really see
vReps as a way to manage customer interactions online, helping not only to
answer questions but also to get info that people might need from a variety
of sources (i.e. Web pages, order tracking and CRM systems, back-end
databases, etc.).
If the vRep cannot answer a customer's question, the virtual service agent
can triage into various channels of support (live chat or call center, for
example) so that the conversation doesn't end there."
Fair enough, and no doubt useful — NativeMinds clients include Ford,
GlaxoSmithKline,
Oracle and Deutsche Telekom, among others.
Ferrara said the company's competition includes Ask Jeeves, which also
markets a natural language technology.
"When it comes to giving customers fast answers, NativeMinds is really
competing with call centers full of live agents on headsets," he said.
My take: We're not there yet, but vReps could be very useful to some
e-commerce sites and we're getting awfully close to the day when you'll be
able to hire a piece of software that's as good as a human being.
By the way, if you are interested in the Turing test, here's a good place get an
education.
In 1991 Dr. Hugh Loebner started the annual Loebner Prize competition — a
$100,000 prize offered to the author of the first computer program to pass
an unrestricted Turing test. When I e-mailed him to ask if the money was ever
claimed, he responded: "It's safe to say they haven't even come close!" The
funds are on deposit with the Cambridge Center and The London Museum of
Science.
Beth Cox has been a well-known keynote speaker and author as well as a business and technology advisor. She helps companies improve their business performance, better utilize data, and understands the implications of new technologies, such as (AI)artificial intelligence, big data, blockchains and the Internet of Things.
